Show Posts

This section allows you to view all posts made by this member. Note that you can only see posts made in areas you currently have access to.


Messages - Dave Yeo

Pages: 1 ... 243 244 [245] 246 247 ... 321
3661
Web applications / Re: Oddities with Youtube.
« on: November 22, 2017, 09:10:38 pm »
The SeaMonkey I'm using. Note that I targeted the Pentium M. I might upload this for experimental purposes.

gcc.exe    4.9.2    -Wall -Wempty-body -Wpointer-to-int-cast -Wsign-compare -Wtype-limits -Wno-unused -Wcast-align -Zomf -std=gnu99 -fgnu89-inline -fno-strict-aliasing -Zomf -fno-math-errno -pthread
c++    4.9.2    -idirafter g:/OS2TK45/h -Wall -Wempty-body -Woverloaded-virtual -Wsign-compare -Wwrite-strings -Wno-invalid-offsetof -Wcast-align -fno-exceptions -fno-strict-aliasing -Zomf -fno-rtti -fno-exceptions -fno-math-errno -std=gnu++0x -pthread -DNDEBUG -DTRIMMED -g -mtune=generic -march=pentium-m -Os -mstackrealign -fomit-frame-pointer

3662
Web applications / Re: Oddities with Youtube.
« on: November 22, 2017, 08:58:47 pm »
Dave,
...About architectures, or at least my understanding. There's two parameters that can be passed to GCC to control what CPU we target.
 -march, which controls the actual instruction set. Targeting the Pentium 4 gives various extra instructions over the i686 such as MMX, SSE and SSE2 and perhaps others. Should benefit all modern CPU's but will sigill (unsupported instruction) on older CPU's
-mtune, which controls the instruction ordering. Most CPU's are fairly happy with the i686 ordering as even on your Phenom II, it should be close to optimum or at least good enough. Then there's the Pentium 4, which liked completely different ordering, ordering that are not optimum for most CPU's due to the P4 using such a different architecture. I'd guess if you look, your Firefox uses something like -mtune=pentium4.

Regarding the Pentium4 and extra instructions, is that really what you meant to say? It is the newer CPUs that introduce additional functionality (SSE2 maybe, etc?) and therefore the Pentium4 would default to a smaller sub-set of instructions, whereas the i686 would be the super-set. So my thinking is that the i686 target is the "desired" code for the newer CPUs...if not, then my thinking is reversed...LOL!

Yes, that is what I meant to say. The i686 is the Pentium Pro, which is older then the Pentium 4 and didn't even have MMX instructions.

Quote
Regarding the -mtune option, only the last 45.9 test build shows the -mtune=i686 config, the other two, the GA 7z and RPM releases do not show the matching Pentium4 nor i686 flags.

That's weird. In some ways tuning for at least an i686 is more important then the CPU instructions.  If nothing else, it has become important to align the instructions on 4 (8?) byte boundaries due to the way all CPU's since the original Pentium access memory.

3663
Web applications / Re: Oddities with Youtube.
« on: November 22, 2017, 06:55:04 pm »
Well, I have firefox-45.9.0-2.oc00.i686.rpm, which is targeted at the i686.  Unluckily it is not yet ready for wide deployment for a couple of reasons. Where to put the icons, and which icons to have. The need to move some stuff from netlabs exp to rel such as libcx and now dmik is working like mad to get the test suite to run so he can search for the causes of instability. He's having problems with the tool chain right now.
I'd assume that both i686 and pentium4 rpms were built and for some reason, only the pentium4 rpm was zipped up and released as a test version.

About architectures, or at least my understanding. There's two parameters that can be passed to GCC to control what CPU we target.
 -march, which controls the actual instruction set. Targeting the Pentium 4 gives various extra instructions over the i686 such as MMX, SSE and SSE2 and perhaps others. Should benefit all modern CPU's but will sigill (unsupported instruction) on older CPU's
-mtune, which controls the instruction ordering. Most CPU's are fairly happy with the i686 ordering as even on your Phenom II, it should be close to optimum or at least good enough. Then there's the Pentium 4, which liked completely different ordering, ordering that are not optimum for most CPU's due to the P4 using such a different architecture. I'd guess if you look, your Firefox uses something like -mtune=pentium4.
I usually build with -mtune=generic, which should be a good compromise for whichever CPU's were common when the version of GCC we're using was released.
Seems your CPU doesn't like the P4 instruction ordering.

3664
Article Discussions / Re: Sponsorship needed for new OS/2 web browser
« on: November 18, 2017, 09:03:14 pm »
The IBM debuggers mostly work fine for GCC(g++) code including Mozilla. Steven did work getting the OpenWatcom linker linking the huge debug version of xul.
I suppose that IBM's HLL symbol format is utilised. That's of course a weak point in the Watcom debugger. So in a real world scenario I have to use a remote debugger under Windows NT or AIX.

Yes, HLL debug symbols are generated with -g -Zomf. Some have to be repressed as there isn't room for the records.
There are plans to add HLL debug symbols to OpenWatcom, though whether time will be found...
Quote

Quote
Firefox itself has tools for debugging JavaScript and such. Profiling with GCC was discussed years back in the newsgroup but I forget the particulars.
That was back in the times when IWB changed to be built with GCC - not sure.
It was around Firefox 3, 9 years back. Here's the discussion, https://groups.google.com/forum/#!searchin/mozilla.dev.ports.os2/GCC$20profiling/mozilla.dev.ports.os2/A9wuXGW8TKA/ZtAu6HerUC4J
Quote

Quote
The newsgroup seems pretty dead.
But is the only source of valuable information that still remains.

Yes, and luckily Google seems to have it all indexed.

3665
Hardware / Re: Hardware Reports / Reviews - More - MORE !!!!
« on: November 18, 2017, 05:25:21 am »
When it comes to USB flash drives, it might be better to collect non-working modules as in my experience, they all work once set up correctly.

3666
Internet / Re: Lightning (SM/TB)
« on: November 18, 2017, 05:23:46 am »
Localization has been fixed, so the next version of Lightning should have a language pack.
Note, it was mostly localization issues, especially on Lightning, that prevented 2.42 and various other versions from being released.
Asking in the newsgroup could be a good idea, as Frank-Rainier Grahl, SM developer, who gave me the patches for 2.42 might have had the correct localized version.

3667
Article Discussions / Re: Sponsorship needed for new OS/2 web browser
« on: November 18, 2017, 03:24:11 am »
It's nice to have GCC (g++) but I cannot find the GNU debugger (gdb)? Which debugger is used for current Mozilla development? Which profilers and code analysers are utilised? Questions, questions that will stay unanswered. The unclear current situation without any plan (as Sigurd mentioned) leads to unsatisfied users etc.

The IBM debuggers mostly work fine for GCC(g++) code including Mozilla. Steven did work getting the OpenWatcom linker linking the huge debug version of xul. Firefox itself has tools for debugging JavaScript and such. Profiling with GCC was discussed years back in the newsgroup but I forget the particulars.

Quote
It would be a better idea to discuss about the technical details in the mozilla OS/2 newsgroup and not here.

The newsgroup seems pretty dead.

3668
Article Discussions / Re: Sponsorship needed for new OS/2 web browser
« on: November 18, 2017, 02:12:22 am »
For the next five to ten years we will still have 32-bit systems (non OS/2 based), including Windows. So it is to be expected that major players will have a 32 bit application. Going to ONLY 64 bit is a big mistake from the Mozilla foundation. Not all can go to 64 bit for several reasons. If we don't have a FF 32 bit then something else will pop up, likely in some kind of public domain license.

Mozilla will likely be the last to drop 32bit support just like they were the last to drop XP and even OS/2.
The problem is that generally developers have the latest machines, using a 64 OS with lots of memory and 32 bit will become more and more of an afterthought. We're seeing it in things like the recommendation to have 16GBs of virtual address space to build Mozilla as one example.
MS will keep supporting 32bit for quite a while, along with some Linux dists but others have already switched to pure 64bit and as programs continue to bloat up, 32 bit will be harder.
There's also more and more dependencies on using the video cards for stuff, and not just drawing. My old gutless cheap phone can do things that I can only dream off in OS/2.

3669
Article Discussions / Re: Sponsorship needed for new OS/2 web browser
« on: November 18, 2017, 02:04:10 am »
The only working approach to ship a Mozilla-forked web browser for oldschool systems is Cameron's TenFourFox http://tenfourfox.blogspot.de/.  So the web browsing experience on a 17 years old G4 is better than on more recent Intel-based hardware under OS/2 nowadays.

Huh? Tenfourfox is stuck on 45ESR and is now forked due to 52ESR not being buildable on PowerPC Macs due to no thread local storage IIRC. Cameron wasn't very happy about it but at least he is keeping certificates, security patches up to date unlike Bitwise as well as other improvements. See the release notes for the latest, https://github.com/classilla/tenfourfox/wiki/ZZFPRReleaseNotes04.
If we had someone with the interest and knowledge, going the same way with our fork would be another possibility. Bitwise has too much on their plate to do much more then basic porting. I don't have the skills and everyone else is busy as far as I know. It's almost a full time job keeping up with Mozilla and ideally we'd need someone with access to the security bugs.

3670
Article Discussions / Re: Sponsorship needed for new OS/2 web browser
« on: November 17, 2017, 11:15:50 pm »
It would be good to continue on to Firefox 52ESR, which should be doable. 52ESR is going to be the last Mozilla release for XP and may be maintained for longer or forked for XP (and Vista) users.
SeaMonkey is likely to die after 52ESR and is currently based on 52ESR with the next few versions planned to stay on 52ESR and there is a good chance that Mozilla52ESR will be forked to enable SeaMonkey to keep working with security fixes, there's rumours that some Linux dists are planning on maintaining it, which would include Firefox. Considering that we're always a year behind, we may be able to use the 52ESR code base for a few years
Thunderbird has also considered forking Mozilla-Central (Firefox head), but that seems unlikely right now and if they do fork it, it'll probably be 56 that is forked.
In summary, there is no great rush to move away from Mozilla and 52ESR may be viable for 3 or 4 years, by which time most browsers are probably going to be 64bit or in the case of QT based webkit, just as out of date as 52ESR. Chromium has the same problems as Mozilla with too fast of an update cycle.
All the other forks of Firefox I've looked at have been 64bit or dependent on working 3D graphics support.


3671
Web applications / Re: Oddities with Youtube.
« on: November 17, 2017, 01:38:42 am »

The CPU load of a slower Pentium 4 desktop often is less than 1% while browsing about:blank. Unfortunately the bad news is that the next game is guessing a frame rate. Any of 6000, 8000, 10240, 12000 and 14000 results in a hanging session (the C-A-D escape always works), a system which isn't responsive, no cursor, or the extremely slow cursor while typing. As if the CPU load is 100%+, while the system displays a CPU load of 0.4%.


It really depends on hardware, though in theory 50-60 should be perfect (does Europe use 50 or 60 hz refresh?)  From what I understand, it was a bad choice on Mozilla's part to use it, even on systems that can read the vsync signal.

I have found the best frame rate to be 59 or 60.

3672
Article Discussions / Re: Sponsorship needed for new OS/2 web browser
« on: November 17, 2017, 01:35:46 am »

But the examples are unfortunately missing here. Speaking about GNU's GCC there's no mention about OS/2 in the current supported releases (6.4 and 7.2).

Our GCC is a fork, forked back in the early 2.x days I believe, possibly due to RMS not wanting anything to do with proprietary systems such as OS/2.
IBM paid for the development of GCC 3.2.2-3.3.5 to replace VACPP rather then develop VACPP anymore, so it is our official compiler I guess.

3673
Web applications / Re: Oddities with Youtube.
« on: November 16, 2017, 05:10:36 am »
Strange that they uploaded the P4 version. As I said, the version I'm running, downloaded from Arca Noae's experimental RPM repository is i686, like all my other RPMs. Unluckily the Arca Noae repository isn't publicly available. I guess have to wait until the build is good enough to post. If I had the bandwidth, I'd post a build.
At least we know why it is crashing :)
As for SeaMonkey, just tested on my T42. Displaying about:addons, with Chatzilla enabled, 90%+ CPU usage. With Chatzilla disabled, 1.2% with the odd peak at 6%. Try it. It does use quite a bit of CPU loading a page but is quite usable.

3674
Article Discussions / Re: Sponsorship needed for new OS/2 web browser
« on: November 16, 2017, 01:48:40 am »
@Martin, it isn't just Rust but lots of other changes. Yesterday they announced that soon they'll require Python3, which we don't have and today, they seem to have made skia a hard dependency, so our Cairo no longer works and skia would need porting. Who knows how many things I've missed as I'm not following development as much as I used to.
They also recommend 16 GBs of address space to compile now, so basically 64 bit platforms. I think that is due to Rust. Eventually we're going to be left behind in a 64bit world

3675
Web applications / Re: Oddities with Youtube.
« on: November 16, 2017, 01:42:27 am »
No(t yet), but I'll see what I can do...

</silence>
My area of expertise: nothing.

Ignoring the additional requirement of the archiver, the name of their file is ... firefox-debuginfo-45.9.0-2.oc00.pentium4.7z.
<silence>

That's weird. Looking at about:buildconfig, they targeted i686 (needed for some atomic instructions, the -march parameter to GCC). Unless you have a different version of mine, an RPM. Check the URL about:buildconfig.
Wish they'd just include the XQS files rather then the huge dbg files.
Unluckily, we stuck having to follow upstream, so it is not Bitwise trying to break old computers, at that previous to 45 they targeted i486 but the build failed with 45 until they upped it to i686, Note I moved to i686 back in the 10ESR days and never had any complaints. Currently Mozilla requires SSE2 support and has since just after 52ESR.
As for SeaMonkey, did you try disabling add-ons? Some, especially chatzilla, cause high CPU usage.
The good news is that dmik is finally getting the tests working and running so hopefully we'll get more stable browser.

Pages: 1 ... 243 244 [245] 246 247 ... 321